Receptionist Agent

Overview

The AI-powered Receptionist Agent provides multi-channel communication, fostering an agentic relationship between patients and healthcare organizations. It facilitates efficient, personalized interactions via SMS, voice calls, and WhatsApp, ensuring a seamless patient experience across all communication platforms.

Top-Level Use Cases

Inbound Call Volume & Outcomes

Healthcare organizations can handle greater call volume with better outcomes and reduced operational costs.

Features:

• Low Latency NLP-based Skill Activation: Loads contextually-specific protocols such as scheduling rules, insurance handling, triage, registration, medications, labs, forms, records, and billing. Includes deferred and warm transfers to appropriate human resources when necessary.

• Low Latency Responses:

  • Simultaneous response planning using partial transcripts within speech loops.

  • Utilization of self-hosted large language models (LLMs).

• Per-Patient Memory:

  • Tracks care goals defined by patients and providers.

  • Multilingual capabilities based on patient language preferences.

  • Saved identity verification and preferred contact methods.

  • Self-scheduled callbacks (e.g., “call me back tonight”).

  • Multi-channel persistence, remembering patient interactions across voice, text, WhatsApp, and more.

Outbound Retention & Registration

Healthcare organizations can increase patient scheduling and registration through automated outreach.

Features:

• Automatically identifies patients requiring follow-up visits by integrating queries from EHRs, waitlists, Sully Scribe integration, and direct API access.

• Dynamically remembers outreach goals and reintroduces them regardless of when conversations occur or who initiates them.

System Architecture

Contextual Integration

The Receptionist Agent utilizes extensive contextual information to deliver personalized and efficient patient interactions:

• Patient Historical Data: Integration of electronic health records (EHR) and Sully Scribe information.

• Dynamic Instruction Loading: Real-time loading of specific instructions based on patient complaints for tailored interactions.

Multi-Channel Capabilities

Seamless operation across SMS, voice, and WhatsApp platforms, ensuring unified communication and continuity of patient experience.

Latency and Voice Optimization

Self-Hosted Infrastructure

Significantly reduces latency, ensuring real-time conversational responsiveness:

• Parallel response generation and pre-generated conversational segments.

Advanced Voice Generation

• Utilizes 11 Labs models for natural, conversational speech.

• Ensures speech normalization and voice-friendly formatting via NLP techniques.

Goal-Oriented Interaction

Proactively manages interactions based on predefined organizational goals:

• Configurable automation workflows such as waitlist management and follow-ups.

• Initiates proactive outbound communications aligned with patient-specific goals.

Warm Transfer Capabilities

Facilitates seamless transitions between automated interactions and human staff, transferring comprehensive patient context for efficient follow-up.

Key Differentiators

• Multi-Domain Functionality: Dynamically adjusts instructional contexts to handle diverse clinical and administrative scenarios.

• Real-Time Adaptability: Adjusts conversational flow and content rapidly, ensuring responsive interactions.

Future Enhancements

Planned improvements include:

• Enhanced Graph-Based EHR Integration: Dynamic EHR graph structure that evolves through interaction-based learning.

• Further Latency Reduction: Continuous optimization of infrastructure for enhanced responsiveness.

• Expanded Channel Integration: Incorporation of additional communication channels to enhance accessibility.

• International Market Expansion: Building dedicated stacks for international customers, including Arabic support for KSA/UAE markets and solutions for Eastern European hospital groups.

  • These implementations will feature high-quality transcription, dialect support, and per-country data isolation to significantly increase our potential market size.

Conclusion

The Receptionist Agent significantly enhances patient interactions through personalized, responsive, and proactive multi-channel communication, ensuring robust performance and streamlining operational workflows.

2 Top-level Use Cases:

• Inbound Call Volume & Outcomes

  • Healthcare organizations can handle greater call volume with better outcomes and lower cost

  • Features:

    • Low latency NLP-based skill activation to load contextually-specific protocols (e.g. scheduling rules for a particular provider)

      • Handle Scheduling, Insurance, Triage, Registration, Medications, Labs, Forms, Records, Billing etc.

      • Defer/Warm transfer to the appropriate resource when the agent can’t handle the request

    • Low latency responses

      • Simultaneous response planning on partial transcript + speech loop

      • Self-hosted LLMs

    • Per-patient memory

      • Remembers goals the patient and their provider have for their care

      • Multilingual, language of preference

      • Saved identity verification + preferred method of contact

      • Self-scheduled callbacks (e.g. “call me back tonight”)

      • Multi-channel (remembers you across voice, text, WhatsApp etc.)

• Outbound Retention & Registration

  • Healthcare organizations can schedule & register more patients with automated outreach

  • Features:

    • Automatically detects patients for follow-up visit scheduling outreach by querying your EHR, integrating with waitlists, integration with the Sully scribe, and direct API access

    • Remembers outreach goals and brings them up dynamically regardless of when the conversation takes place or who started it